Personifying the radical: How news framing polarizes security concerns and tolerance judgments.

Gil de

Zuniga

- Edu

catio

nal O

nly

Keum et al. / FRAMING CIVIL LIBERTIES 337

Human Communication Research, Vol. 31 No. 3, July 2005 337–364© 2005 International Communication Association

Personifying the RadicalHow News Framing Polarizes Security Concerns and Tolerance Judgments

HEEJO KEUMUniversity of Texas at San AntonioELLIOTT D. HILLBACKHERNANDO ROJASHOMERO GIL DE ZUNIGADHAVAN V. SHAH DOUGLAS M. MCLEODUniversity of Wisconsin-Madison

This study examines relationships among individual dispositions, news framing of civil lib-erties restrictions, security concerns, and political tolerance. We theorize that news frames condition the effects of individual dispositions on security and tolerance attitudes. To explore these relationships, an online-survey experiment was conducted with 650 respondents. This experiment presented alternative versions of news stories about domestic security policies following September 11, and the policies’ implications for a fringe activist group. One factor was whether the activists targeted by the government advocated for a cause supported or opposed by the respondent; another factor was whether the story framed government actions against the activists at the individual or group level. Findings show that individual fram-ing—as opposed to group framing—made participants less tolerant of radicals they opposed and more tolerant of radicals they supported. Similar effects were observed for political ideol-ogy. Implications of personifi cation as a framing device are discussed.

The terrorist attacks of September 11, 2001, heightened a long-standing clash between two fundamental democratic values: the protection of public safety and the preservation of civil liberties. In

the wake of these events, the U.S. Congress passed the USA PATRIOT

Heejo Keum (Ph.D., University of Wisconsin-Madison) is an assistant professor in the De-partment of Communication at the University of Texas at San Antonio. Elliott Hillback (M.A., University of Wisconsin-Madison) and Homero Gil de Zuniga (M.A., University of Wisconsin-Madison) are doctoral students in the School of Journalism and Mass Com-munication at the University of Wisconsin-Madison. Hernando Rojas (Ph.D., University of Wisconsin-Madison) is an assistant professor in the Department of Life Sciences Com-munication at the University of Wisconsin-Madison. Dhavan V. Shah (Ph.D., University of Minnesota) is a professor in the School of Journalism and Mass Communication and the Department of Political Science at the University of Wisconsin-Madison. Douglas M. McLeod (Ph.D., University of Minnesota) is a professor and head of graduate studies in the School of Journalism and Mass Communication at the University of Wisconsin-Madison. Correspondence concerning this article should be addressed to Heejo Keum, Department of Communication, University of Texas at San Antonio, 6900 N. Loop 1604 W., San Antonio, TX 78249–0643; email: [email protected].

Gil de

Zuniga

- Edu

catio

nal O

nly

338 HUMAN COMMUNICATION RESEARCH / July 2005

(Uniting and Strengthening America by Providing Appropriate Tools Required to Intercept and Obstruct Terrorism) Act. Some people wel-comed its passage as a deterrent to future threats. The vague language and considerable reach of this act, however, has generated concern that it could be used to infringe on privacy and due process rights of political radicals—activists who may stake out extreme, often unpopular, posi-tions on issues but are not suspected of criminal activities. These fears have been realized, with the U.S. Justice Department acknowledging that the act has been used to infi ltrate and investigate nonviolent protest groups (Goldberg, 2004).

We argue that the nature of news coverage about civil liberties con-troversies has the potential to sway individuals’ security concerns and social tolerance judgments. Our theory builds on past research that fi nds contemporary information works in combination with citizens’ politi-cal predispositions to shape the level of support for civil liberties (Mar-cus, Sullivan, Theiss-Morse, & Wood, 1995). This research attempts to explain how individuals respond when confronted with disliked groups; we argue that it is under these conditions that the limits of tolerance are best understood (Sullivan & Marcus, 1988; Sullivan, Piereson, & Marcus, 1979). An unintended consequence of this focus on least-liked groups has been a dearth of research on how individuals make these judgments when they confront efforts to restrict the civil liberties of groups whose causes they are predisposed to support—but whose tactics they may op-pose. The decision to defend civil liberties of radicals, even if only for targets towards which one feels some latent sympathy, is a meaningful test of tolerance in the current political climate.

More important, previous empirical research has studied tolerance toward groups, even though the individual is the traditional center of democratic governance stressing civil liberties (Golebiowska, 1996). In-deed, individual liberties are privileged in the Bill of Rights and other constitutional systems. The need to consider how a focus on individuals rather than groups shapes tolerance judgments is further highlighted by reporters’ tendency to personalize press reports through episodic story frames (Iyengar, 1991). This journalistic propensity to personify the news has been found to condition media effects and shape cognitive processes (Shah, Kwak, Schmierbach, & Zubric, 2004). Building upon these insights, this study expands the study of framing effects on support for civil liber-ties. We consider whether news framing in individual and groups terms differentially shapes security concerns and tolerance judgments when people encounter restrictions on fringe activists they are predisposed to support or oppose.

Gil de

Zuniga

- Edu

catio

nal O

nly


THEORETICAL FRAMEWORK

Insight into the infl uence of contemporary information on tolerance judgments may be gained from three domains of scholarship: (a) political theory and research on the democratic foundations of political tolerance; (b) work on the causes and consequences of support for civil liberties; and (c) research on news framing as it relates to biases in social attribu-tion and impression formation processes. We integrate these literatures for our theory of how individual news frames polarize security concerns and tolerance judgments.

Political Tolerance in Democratic Context

In the United States, the preservation of civil liberties and the toler-ance of differences have historically been favored over public safety and national security. The ideology of “classic” liberalism encouraged the framers of the Constitution to privilege individual rights, and to guard against the state’s potential power to infringe on them. Even with this emphasis on the rights-bearing individual, concerns about safety and the potential for harm have been cited to justify policies that run counter to these guarantees. Indeed, security concerns have sometimes trumped civil rights during periods of crisis, most notably in the McCarthy Era. During that time, apprehension about a fi fth column of internal threats was used to justify incursions into the private sphere. Some observers have compared the current “War on Terror” to these Cold War practices, noting that concerns about security and liberty once again defi ne citi-zens’ tolerance judgments in the face of growing incursions by govern-ment into the private sphere.

Since Samuel Stouffer’s (1955/1963) classic analysis of attitudes to-ward nonconformists, numerous studies in political science have exam-ined how citizens navigate this value confl ict and form tolerance judg-ments. In the 1950s, Stouffer was concerned with two threats to Ameri-can democracy: a “communist conspiracy” and “those who in thwarting the conspiracy would sacrifi ce some of the very liberties which the en-emy would destroy” (p. 13). Stouffer’s research tried to establish empiri-cally how liberties could be preserved, despite the presence of perceived threats. To him, intolerance was a function of lower levels of education and higher levels of perceived threat. Ideology was indirectly implicated in this dynamic.

Other research has related tolerance to liberal ideology (Bobo & Licari, 1989; Green & Guth, 1991; Lipset & Raab, 1970). The concept of political ideology assumes polarization of opinions along the liberal–conservative

Gil de

Zuniga

- Edu

catio

nal O

nly


continuum, with liberals more likely to be tolerant and conservatives more wary of nonconformists (Lipset & Raab, 1970; Sullivan, Marcus, Feldman, & Piereson, 1981). Political ideology must be conceptualized with care, however, making special effort to “distinguish between eco-nomic and non-economic issues, where the former refers to questions of the distribution of wealth and the latter to those of cultural conformity and nonconformity” (Sullivan et al., 1981, p. 97). That is, social ideology shapes tolerance.

This research defi ned tolerance as an attitude “that allows people to have freedom of expression even though one may feel their ideas are in-correct or even immoral” (Nunn, Crockett, & Williams, 1978, p.12). Even though support for such freedoms rose in the wake of the McCarthy Era, Nunn and colleagues (1978) cautioned against being overly optimistic: “Given a realistic menace to domestic tranquility enduring over a length of time, the swelling upward trend in support of civil liberties toward which most signs now point might well subside” (p. 95). Their data show that intolerance rose for select groups even as it was declining toward communists.

Reconceptualizing Political Tolerance

Expanding upon these insights, Sullivan and colleagues (Sullivan, et al., 1981; Sullivan, Piereson, & Marcus, 1979) responded to the fact that previous studies typically measured intolerance toward leftist groups, not general intolerance. They conceived of tolerance as “a willingness to ‘put up with’ groups that people may oppose or dislike” (p. 249) and considered people to be tolerant only if they extend civil liberties to groups they were predisposed to dislike. Testing this conceptualization, Sullivan, Piereson, and Marcus (1982) found that the educated were not much more tolerant than the uneducated, largely eliminating the appar-ent gains in support for civil liberties. The overestimation of education’s impact on tolerance stemmed from the fact that the least-liked groups of educated people tend to be on the political right, whereas leftist groups were typically least liked by less educated people.

Marcus et al. (1995) found that when people were exposed to disliked groups in experimental settings, three sets of factors shaped their toler-ance judgments: predispositions (threat predisposition, neuroticism, and gender); standing decisions (tolerance of least-liked groups, support for democratic principles, and political knowledge); and contemporary infl u-ences (appraisals of threat and their impact on processing strategies). In line with their conceptualization, they tested the limits of tolerance by gauging subjects’ predispositions and then presenting them with infor-mation about groups they were liable to fi nd objectionable. This research

Gil de

Zuniga

- Edu

catio

nal O

nly


technique has considerably expanded the range of targets examined by scholars interested in public support for civil liberties and the factors underlying tolerance judgments.

The focus on least-liked groups, however, has also limited understand-ing of public responses to incursions on the civil liberties of radicals and nonconformists who advocate in behalf of supported causes. As a result, an important question remains unanswered: How do citizens respond to threats against the civil liberties of radical groups that endorse causes they favor rather than oppose? As Chong (1994) noted, a focus on oppo-sitional groups has limited the scope of research on civil liberties judg-ments and the understanding of processes underlying these evaluations and decisions. Measures of tolerance, he argued, should not concentrate exclusively on disliked groups; they should also “discover the range of groups, ideas, and activities that people will allow to go uncensored in society, and the circumstances in which such tolerance will be given” (p. 27). Research on tolerance has benefi ted from a focus on disliked groups; however, it needs to consider how citizens respond to government ef-forts to restrict the speech freedoms and privacy of less oppositional, but nonetheless controversial, targets.

This is particularly true in the wake of the September 11 attacks and the start of the War on Terrorism—for, as we contend, concern about the effect of a realistic, long-term threat to domestic security is again acutely relevant. Given the current policy and opinion climate, it becomes nec-essary to consider not only tolerance for disliked radicals, but also the willingness of individuals to defend the civil liberties of political radi-cals with whom they share a common cause. This orientation toward a broader range of civil liberties targets refl ects the belief that the degree to which people are willing to extend civil liberties to disliked social groups is only one point on a continuum of tolerance judgments.

Further, the decision to defend civil liberties, if even only for selected radicals, is a particularly meaningful test of the trade off between civil liberties and national security in the current policy and opinion climate. A willingness to restrict the liberties of extremist activists whose ideol-ogy one is predisposed to support may refl ect a more general orienta-tion toward supporting government efforts to combat terrorism, broadly defi ned. Conversely, opposition to these efforts suggests that security concerns may be trumped by the desire to protect the civil liberties of certain types of radical elements. By examining respondents’ attitudes toward militant activists that they are predisposed to support or op-pose, it is possible to see how orientations toward targeted groups affect judgments about civil liberties and security. In other words, tolerance judgments that balance civil liberties and national security concerns are likely to vary according to an individual’s orientation toward the tar-

Gil de

Zuniga

- Edu

catio

nal O

nly


geted group. As a result, in addition to Sullivan and colleagues’ focus on disliked groups, an examination of citizens’ security concerns and tolerance judgments in contemporary America necessitates some con-sideration of how people respond to restrictions on the civic liberties of radicals, whose cause they may be predisposed to support but whose tactics they may fi nd distasteful.

News Framing and Impression Formation

Another defi ciency of contemporary research is its inattention to in-dividuals as the object of tolerance judgments. This is particularly trou-bling given the focus on the individual in classical democratic theory. Consistent with this concern, Golebiowska (1996) found that information about the object of toleration infl uenced public evaluations, underlin-ing the need to understand how evaluations of individuals and groups shape these judgments. Research on perceptions of individuals and per-ceptions of groups has generally followed separate lines, with few di-rect comparisons being made between them (see Hamilton & Sherman, 1996). The effects of framing styles on social judgments can be better understood by juxtaposing the two.

One domain of inquiry that does consider how people respond dif-ferently to comparable information about individuals and groups is re-search on news framing, social attribution, and impression formation. This work begins with the assumption that by framing social and po-litical issues in specifi c ways, media imply the underlying causes and consequences of a problem and provide standards on which evaluations and judgments can be based (Iyengar, 1991; Shah, Domke, & Wackman, 1996). Most studies assert that frames trigger a set of mental processes (i.e., the application of certain heuristics or the exploitation of cognitive biases) and thereby encourage ways of understanding and shape judg-ments about issues (Price & Tewksbury, 1997).

Research examining framing effects on tolerance judgments has found that, when news coverage highlights civil liberties over safety risks, those exposed to it are more socially tolerant (Nelson, Clawson, & Oxley, 1997). Journalistic norms and news standards nonetheless dictate that press reports rarely feature a single organizing value; instead, they tend to present controversial issues as “value-choices,” stressing the confl ict between contending positions (Ball-Rokeach & Loges, 1996; Shah et al., 1996). Examples of this may be found in news coverage of protestors and radicals, where stories often highlight the confrontation between their positions and the status quo. Research on the journalistic “protest para-digm” indicates that radical groups are rarely treated as equal partners in such confl icts. Through various conventions such as focusing on events

Gil de

Zuniga

- Edu

catio

nal O

nly


rather than underlying issues, and featuring the most radical individuals in representations of protest groups, journalists tend to marginalize pro-test groups (McLeod & Hertog, 1999).

Similar observations have been recognized in the news framing lit-erature, which notes that journalists tend to frame their reports around particular instances and individual exemplars (Iyengar, 1991). The news media’s tendency to favor episodic frames (typically focused on discrete events and their infl uence on individuals) over thematic frames (typi-cally focused on social trends and their effects on communities) sig-nifi cantly affects how viewers attribute responsibility for solving social problems. Episodic framing is thought to foster “attributions of respon-sibility both for the creation of problems or situations (causal responsi-bility) and for the resolution of these problems or situations (treatment responsibility)” (Iyengar, 1991, p. 3). By contrast, framing the issue in more general terms is theorized to encourage a sense of shared respon-sibility for addressing the social issue. Experimental studies involving crime, poverty, and unemployment provide limited, largely nonsignifi -cant, support for these claims.

These weak fi ndings may be explained, in part, by recent work that asserts Iyengar’s distinction between episodic and thematic framing may confound two frame dimensions (Shah et al., 2004). Episodic coverage not only favors specifi c instances over enduring problems (i.e., time span), it also emphasizes individual situations over societal conditions (i.e., social level). This second factor, the distinction between individuals and collec-tives, is particularly relevant for tolerance judgments given what we know about the social cognitive processes. In particular, research on impression formation draws a distinction between judgments about individuals and groups that is directly relevant to this journalistic norm.

This work emphasizes the signifi cance of a continuum of perceived entitativity in making social judgments (Hamilton & Sherman, 1996). En-titativity—the degree to which social units are perceived as being cohe-sive and consistent—ranges from very high in individuals, to less so in groups, to lowest in population aggregates. When subjects perceive a high degree of entitativity, as when perceiving an individual, they ex-pect a greater degree of consistency in attitude and behavior. As such, subjects feel more comfortable in making trait inferences. When making judgments about a group, however, a lower perceived entitativity makes subjects less likely to make such inferences, instead basing their percep-tions on a wider range of criteria.

The overall result, note Hamilton and Sherman (1996), is that sub-jects are quicker to make trait inferences about individuals than about groups, and that these inferences are more pronounced. Consistent with this work, Susskind et al. empirically fi nd that “in comparison with those

Gil de

Zuniga

- Edu

catio

nal O

nly


forming impressions of a group or an aggregate, participants in in-dividual target conditions made stronger (more extreme) trait judg-ments, they made those judgments faster, and they made them with greater confi dence” (Susskind, Maurer, Thakkar, Hamilton, & Sher-man, 1999, p.190).

HYPOTHESES

Role of Cause Predisposition

Scholars agree that negative evaluations of a group play a signifi cant role in tolerance judgments (Kuklinski, Riggle, Ottati, Schwarz, & Wyer, 1991; Marcus et al., 1995; Hurwitz & Mondak, 2002). If individuals dis-like a category of radicals or nonconformists, they will judge the ac-tions of those people to be undesirable, threatening, and therefore less worthy of toleration (Marcus et al., 1995; Hurwitz & Mondak, 2002). If, however, individuals encounter government efforts to infringe the liberties of radicals with whom they share support for a cause, they are likely to form tolerant attitudes and maintain members’ civil lib-erties. Under the current political context, both of these evaluations are important for generating a broader understanding of the processes underlying judgments that force trade-offs between civil liberties and national security.

Our research manipulated whether respondents were exposed to a so-cially radical target of civil liberties restrictions that either aligned or did not align with the respondents’ political predispositions. Even though re-search subjects had not encountered this specifi c target before, the politi-cal cause the target represented was familiar to them in that subjects pre-viously indicated that they supported or opposed this cause. We predict-ed that, even when people encounter an unfamiliar group that advocates for a supported cause, they identify with the group and feel motivated to support its civil liberties. Even when a specifi c allegiance does not exist, the mere knowledge of a group’s political orientation is enough to pro-duce a sense of affi liation or opposition (Tajfel, Billig, Bundy, & Flarnent, 1971; Billig & Tajfel, 1973; Tajfel & Billig, 1974; Tajfel & Turner, 1979). This work, under the rubric of the minimal group paradigm, relates this dif-ference to social identity needs (Turner, Hogg, Oakes, Reicher, & Wether-ell, 1987) and reciprocity expectations (Rabbie, Schot, & Visser, 1989; Ng, 1981; Gaertner & Insko, 2001). Thus, support for expressive rights should be reduced and security concerns heightened when people encounter radicals who espouse opposed views. These effects should be reduced, though not necessarily reversed, when people encounter radicals who endorse supported views. Accordingly, we hypothesized:

Gil de

Zuniga

- Edu

catio

nal O

nly


H1a: People who encounter civil liberties restrictions on radicals they are pre-disposed to oppose will have greater security concerns than people who encounter civil liberties restrictions on radicals they are predisposed to support.

H1b: People who encounter civil liberties restrictions on radicals they are pre-disposed to oppose will be less tolerant of expressive rights than people who encounter civil liberties restrictions on radicals they are predisposed to support, even after controlling for security concerns.

Role of Political Ideology

We also hypothesized concerning the effects of political ideology (i.e., liberal versus conservative) on security concerns and tolerance judg-ments. Many scholars assert that the public’s attitude toward these issues stems from fundamental ideological beliefs. Marcus, Sullivan and col-leagues question whether liberals are more tolerant than conservatives when confronted with disliked groups, although others have shown that certain features of liberal ideology promote tolerance and openness to-ward diverse beliefs (Tetlock, 1986; Tetlock, Armor & Peterson, 1994). In fact, empirical studies have confi rmed a relationship between liberal so-cial ideology and tolerance (Lipset & Raab, 1970). Thus, we predicted that people who endorse a liberal social ideology will show greater support for civil liberties and hold fewer security concerns. We hypothesized:

H2a: People who are ideologically liberal will have lower security concerns than people who are ideologically conservative.

H2b: People who are ideologically liberal will be more tolerant of expressive rights than people who are ideologically conservative, even after control-ling for security concerns.

Role of News Framing

Further, we expected that the main effects of support for or opposi-tion to a cause on these judgments will be different, depending on the social level at which the news story is framed. Work on social perception provides a basis for this expectation. Research on social attributions and impression formation fi nds that perceivers make more extreme judg-ments for an individual than for a group, because they view an individ-ual as a coherent and unifi ed psychological unit (Hamilton & Sherman, 1996; Susskind et al., 1999). In social judgments about individuals under government scrutiny, perceivers assume great consistency between be-liefs and actions, therefore assuming the best or the worst about these individuals (Susskind et al., 1999). By contrast, social judgments are less likely to depend on such extreme inferences when the target is a group,

Gil de

Zuniga

- Edu

catio

nal O

nly


because evaluations of collectives are tempered (Iyengar, 1991; Shah et al., 2004).

Relying on this argument, we hypothesized that the individual versus group framing conditions will amplify or attenuate, respectively, the ef-fects of cause predisposition. We hypothesized that stories focusing on an individual rather than a group will enhance the infl uence of predis-position to oppose or support a cause on security concerns and tolerance judgments. Those who encounter the group framing of this controversy are expected to form evaluations that are less extreme—whether favor-able or unfavorable—with regard to concerns about public safety or civil liberties. Supporting this view, Susskind et al. (1999) observe that toler-ance judgments about individual targets would be more extreme than the judgments about group targets.1 Thus, we predicted that the indi-vidual story frame will be more likely to polarize security concerns and tolerance judgments between those who support and those who oppose the cause.

Likewise, attribution theory suggests that news consumers who re-ceive a story about an individual targeted by the government are more prone to dismiss these civil liberties encroachments and become increas-ingly intolerant, attributing responsibility for the problem to the individ-ual (see Iyengar, 1991). Conversely, when people encounter a story about an individual engaged in behavior they deem positive or favorable, they are likely to empathize with the subject of the story, perceive less of a threat, and be more supportive of civil liberties. Thus, we expected that the individual story frame will condition the effects of cause predisposi-tion, such that security concerns and tolerance judgments are polarized under this condition.

H3a: The infl uence of cause predisposition (an individual’s predisposition to support or oppose the radicals’ cause) on security concerns will be ampli-fi ed when participants encounter a news story framed around restrictions of an individual’s civil liberties, as opposed to one that is framed around restrictions of a group’s civil liberties. That is, news framing will moderate the effects of cause predisposition on security judgments.

H3b: The infl uence of cause predisposition (an individual’s predisposition to support or oppose the radicals’ cause) on tolerance judgments will be amplifi ed when participants encounter a news story framed around re-strictions of an individual’s civil liberties as opposed to one that is framed around restrictions of a group’s civil liberties. That is, news framing will moderate the effects of cause predisposition on tolerance judgments.

Finally, to test further the interplay of frames with political predisposi-tions known to infl uence judgments about safety and liberty, we hypoth-esized that political ideology interacts with individual and group framing

Gil de

Zuniga

- Edu

catio

nal O

nly


of this controversy to shape tolerance and security judgments. Just as individual and group frames are expected to amplify or attenuate the ef-fects of cause predispositions, we believed the same underlying process of entitativity should condition the effects of political ideology on these political evaluations. That is, individual framing is expected to amplify individuals preexisting political leanings such that liberals become less concerned about security issues and more likely to support civil liberties when they encounter individually framed news accounts of government restrictions. Likewise, conservatives become more concerned about se-curity and less likely to support civil liberties when they encounter such news accounts. Thus, we posed the following hypotheses:

H4a: The infl uence of political ideology on security concerns will be amplifi ed when participants encounter a news story framed around restrictions of an individual’s civil liberties as opposed to one that is framed around restric-tions of a group’s civil liberties. That is, news framing will moderate the effects of political ideology on security judgments.

H4b: The infl uence of political ideology on tolerance judgments will be ampli-fi ed when participants encounter a news story framed around restrictions of an individual’s civil liberties as opposed to one that is framed around restrictions of a group’s civil liberties. That is, news framing will moderate the effects of political ideology on tolerance judgments.

METHODS

Design

This study employs a 2 X 2 X 2 experiment embedded within a Web-based survey. Respondents for the survey were drawn from two popu-lations. The fi rst was a group of students enrolled in courses at a large midwestern U.S. university; they were offered extra credit by their in-structors for participating in this research experience. The second group of subjects was recruited from Dane County, Wisconsin residents using a probability sampling variation of random-digit dialing; they were of-fered a chance to win one of fi ve $50 cash prizes in exchange for their in-volvement in the study. Students in an upper-division research methods course contacted sampled households and spoke to a randomly selected adult within each household. Notably, 37.2 percent of the eligible adults who were contacted provided an email address.2

All potential participants were contacted by email and given the Web site of the online survey. Nonstudent participants who provided an in-valid email were recontacted by telephone, and every effort was made to obtain a legitimate email address to use for recruitment. In order to re-ceive extra credit or be included in the prize drawing, participants were

Gil de

Zuniga

- Edu

catio

nal O

nly


required to provide their email at the start of the survey. Individuals who failed to respond to our initial email contact received an email reminder. Approximately 65% of recruited students completed the survey (N = 413). In addition, 51% of nonstudent respondents who provided email addresses took the survey (N = 237). The combined N was 650.

In addition to a standard battery of pretest and posttest questions, the respondents read and responded to an experimentally manipulated, fi ctional news story about potential civil liberties restrictions follow-ing September 11, 2001. This story appeared in several sections. The main section introduced an individual or group targeted by the FBI for monitoring because of an unspecifi ed possible threat. At the end of this brief story, respondents had four choices. They could continue with the survey, or they could read more information in one of three categories: tracking and monitoring, search and seizure, or restrictions on speech and assembly. Each section contained information about additional FBI efforts regarding the target individual or group. At the end of each sec-tion, individuals could either read more about that topic or switch to a different topic. Each topic had three levels of content, meaning that re-spondents could read up to nine additional story segments (beyond the main section) if they chose to do so.

Embedded within these stories were two manipulations that are the focus of this study, as well as a third that is not central to our analysis. The fi rst concerned whether the radical activist targeted by the govern-ment represented a cause that the respondent supported or opposed. This was determined by previously stated preferences to a list of es-tablished special interest groups. Early in the survey, respondents were presented with six mainstream special interest groups and were asked toward which one they felt most negative (see Appendix A for the exact question wording and the names of the six groups).3 A second question asked respondents toward which group they felt the most positive. The answers to these questions were used to generate the cause for which the individual or group featured in the news story was advocating, with respondents randomly assigned to read about a target that they were predisposed to support or oppose. To minimize the infl uence of exist-ing knowledge about real groups and to help mask this manipulation, the group mentioned in the story was an “extremist” offshoot group with similar beliefs to the existing special interest group. For example, if respondents listed Greenpeace as their most-liked group and were as-signed to the predisposition-to-support condition, they would read a story that mentioned the more radical “Earth Defense Front.”

The second experimental factor dealt with how the story was framed. In the group condition, respondents read a story in which the selected radical activist group was the subject of FBI scrutiny. The group was dis-cussed as a unit, and any quotes came from an anonymous spokesperson

Gil de

Zuniga

- Edu

catio

nal O

nly


for the group. Where possible, the story made reference to “groups” rath-er than “individuals.” In the individual condition, subjects read about a particular member of the selected group, Greg Anderson. Anderson was quoted and the story made frequent reference to “individuals” rather than “groups.”

A third experimental factor concerned the ideological cue attached to a fi ctional think tank, the Liberty Institute. Throughout the story the in-stitute was used to provide a counter to FBI claims, discussing the merits of civil liberties protections and the need for government restraint. As a cue to subjects, the institute was described at multiple points in the story as “liberal” for one condition and “conservative” for the other condi-tion. This manipulation is not the focus of this study and was therefore controlled in all tests of hypothesized relationships. Also, to ensure that it did not condition the effects of either of the other two experimental factors, tests were run to confi rm an absence of interactions. Notably, no interactions were observed between this factor and the other two ma-nipulations, so it was simply included as a covariate.

Measures

Criterion Variables

After reading the manipulation stories, subjects answered questions focusing on the criterion variables: security concerns and tolerance for the targeted group. For security concerns, an index was created by aver-aging the scores from two items asking subjects how concerned they are about their own safety and the safety of other people in the U.S. from vi-olent acts by extremists (see Appendix A). Subjects rated their agreement with the statements using a 10-point scale, ranging from not at all con-cerned to extremely concerned (Cronbach’s α = .80; M = 5.16; SD = 2.42).

Tolerance for the targeted group was operationalized with an additive index of four statements taken from Marcus et al. (1995) but modifi ed to fi t the current social context. Subjects were asked how they felt about a set of statements regarding the treatment of the hypothetical group that had appeared on the manipulation stories: Group members should be allowed to work as a teacher in public schools, hold public rallies, broadcast public access cable programs, and share their views over the Internet. Items were measured on 10-point scale from strongly disagree to strongly agree. We used responses to create an index averaging the scores from these items (Cronbach’s α = .77; M = 7.12; SD = 1.86).

Between-Subjects Factor

In addition to the experimental factors—cause predisposition and group framing—we included individual ideology as a between-subjects

Gil de

Zuniga

- Edu

catio

nal O

nly


factor. Consistent with past research, ideology was tapped using a 7-point scale ranging from very liberal to very conservative. Subjects were asked to locate themselves in terms of their position on social issues and people’s behavior (M = 3.08; SD = 1.60).

Blocking Variables

Even in an experimental design, controls help ensure that an experi-mental condition is not confounded with other factors in accounting for criterion variables (Keppel, 1991; Kirk, 1995). To exclude potential con-founds, we blocked a range of potential infl uences on tolerance: demo-graphics, orientational–attitudinal variables, communication patterns, and the other experimental condition.

Previous researchers have found that political tolerance tends to be as-sociated with basic demographic variables (Marcus et al., 1995; Stouffer, 1955, 1963; Sullivan et al., 1981). They demonstrated that female gen-der, younger age, and high social status were likely to be linked to in-creased tolerance. In this paper, a standard set of four basic demographic variables was included: gender (female = 63.1%), age (M = 27.33; SD = 12.69), household income (Mdn = $30,000–50,000), and education (M = 14.80; SD = 3.54). Household income was assessed on a six-point scale (see Appendix A) and education was measured by asking respondents the highest year of school completed. In addition to the demographic controls, this study includes orientation variables tapping materalism and interpersonal trust related to different levels of tolerance. As dis-positional factors, values constitute a basis for political thinking and attitudes (Ball-Rokeach & Loges, 1994; Kinder & Sears, 1985; Krosnick, Berent, & Boninger, 1994; Sotirovic & McLeod, 2001). In particular, re-cent research has highlighted the connection between nurturing ma-ternalism and social tolerance (Lakoff, 2004). Sullivan and colleagues (1981) demonstrated that people who trust others displayed higher levels of political tolerance. In this study, maternalism was measured with two items on a 10-point scale asking the respondent a sense of superiority and benefi cence (Cronbach’s α = .76; M = 7.65; SD = 1.74). Interpersonal trust was also assessed with two items (Cronbach’s α = .54; M = 5.74; SD = 1.66).4

Three communication variables—newspaper hard news use, televi-sion hard news use, and interpersonal discussion of political issues—were also considered to be potential confounds for our study of political tolerance. Even though there has not been much research directly ad-dressing the relationship between communication and tolerance, news consumption and interpersonal discussion are thought to promote politi-cal tolerance (Sullivan et al., 1981). Through various communication chan-nels, citizens can be exposed to diverse people and learn the democratic

Gil de

Zuniga

- Edu

catio

nal O

nly


principle that “free exchange of ideas is necessary and that to be different is not necessarily to be dangerous” (Sullivan et al., 1981, p. 94).

Our measure of newspaper hard news use consists of four measures of attention and exposure to hard news content about international affairs, national politics, local government, and community issues (Cronbach’s α = .79; M = 5.79; SD = 2.27). Television hard news use was also con-structed from four measures of exposure and attention to content about international affairs, national politics, local government, and community issues (Cronbach’s α = .86; M = 5.67; SD = 2.12). Exposure and attention were measured on 11-point scales ranging from never to frequently, and from none to very close attention, respectively. Interpersonal discussion is summed from four measures of frequency of discussion about politics and current affairs (Cronbach’s α = .77; M = 5.72; SD = 1.96). The groups are family members, friends, neighbors, and co-workers. We measured frequency of discussion on a 10-point scale.

Finally, we accounted for the source cue that the subject encountered in the manipulation story and the number of manipulation pages they visited during the experiment. As we noted above, in our experiment, subjects could choose to read up to nine additional story segments (be-yond the main section). The focus of this study is on framing and group like–dislike; accordingly, these study design factors were used as control variables.

RESULTS

To test the hypothesized relationships—the main effects of cause predisposition and ideology, as well as their interactions with the news frame—we conducted analyses of covariance (ANCOVA).5 Demograph-ics (gender, age, income, and education), orientations (maternalism and interpersonal trust), communication (newspaper hard news, television hard news, and interpersonal discussion), and study design elements (source cue and number of pages visited) were included as covariates in the analyses, functioning as blocking variables.

Security Concerns

To simultaneously examine hypotheses related to security concerns, we ran an ANCOVA with the main effects, interactions, and covariates noted above (see Appendix B for full table). This ANCOVA included specifi ed contrast tests in which critical differences between different in-dividual and group frame conditions were gauged. Main effects of cause predisposition and ideology on security concerns were found to be sig-nifi cant (cause predisposition: F(1, 576) = 11.76, p < .001, ηp

2 = .02; ideol-ogy: F(1, 576) = 6.03, p < .05, ηp

2 = .01). As indicated in Table 1, nonliberals

Gil de

Zuniga

- Edu

catio

nal O

nly


and respondents who encountered activists they were predisposed to oppose were more likely to be concerned about security. Thus, H1a and H2a were supported.

H3a and H4a posited that an individual news frame would amplify the effects of cause predisposition and political ideology, respectively, on security concerns. As expected, cause predisposition interacted with the news frame on security concerns, F(1, 576) = 5.23, p < .05, ηp

2 = .01. In contrast, ideology did not have signifi cant interactive effects with the news frame on security concerns, F(1, 576) = .12, n.s., ηp

2 = .006. These fi ndings provide some preliminary support for H3a, yet offer little sup-port for H4a.

To decompose the news frame–cause predisposition interaction, we ran two sets of contrast tests: (a) one set to examine whether there were signifi cant differences between the predisposed-to-oppose condition and the predisposed-to-support condition among respondents receiving the same news frame and (b) another set to examine whether there were signifi cant differences between the individual and group frame condi-tions among respondents sharing the same cause predisposition. The fi rst set provides a limited test of whether the effects of cause predisposi-tions are amplifi ed under the individual frame condition and attenuated

Predisposition Support 4.61 5.13 4.87 7.52 7.23 7.37

(.191) (.200) (.137) (.136) (.139) (.097) N = 151 N = 146 N = 297 N = 151 N = 144 N = 295

Oppose 5.70 5.35 5.52 ` 6.48 6.99 6.73 (.188) (.200) (.137) (.13) (.142) (.098) N = 156 N = 140 N = 296 N = 156 N = 139 N = 295

IdeologyLiberal 4.87 5.02 4.95 7.41 7.25 7.33

(.166) (.171) (.120) (.118) (.122) (.085) N = 196 N = 182 N = 378 N = 196 N = 180 N = 376

Other 5.44 5.46 5.45 6.59 6.97 6.78 (.221) (.229) (.160) (.157) (.163) (.144) N = 111 N = 104 N = 215 N = 111 N = 103 N = 214

TABLE 1Estimated Means for Security Concerns and Tolerance Judgments by Condition

Security concerns Group tolerance

Frame Individual Group Total Individual Group Total

NOTE: Entries in each cell refer to estimated means after accounting for blocking variables. Standard errors are noted in parentheses.

Gil de

Zuniga

- Edu

catio

nal O

nly


under the group frame condition. The second, more conservative, set of analyses provide a test of whether the effects of individual and group news frames are signifi cantly different from one another within cause predisposition conditions and if these differences are in the expected direction.

Consistent with expectations, there was a signifi cant difference in the security concerns of respondents in the predisposed-to-support and the predisposed-to-oppose conditions among respondents receiving the in-dividual news frame (contrast estimate = -1.093; s.e. = .264; p = .000), but not among those receiving the group news frame (contrast estimate = -.214; s.e. = .275; p = 436). This set of fi ndings confi rms that it is under the individual frame condition that the effects of cause predispositions on security concerns are heightened, whereas these same differences are attenuated under the group frame condition. Further, among those who were predisposed to support the target, respondents receiving the indi-vidual frame condition had signifi cantly lower security concerns than people receiving the group frame condition (contrast estimate = -.538; s.e. = .269; p = .046). Even though the general pattern of amplifi cation was maintained among those predisposed to oppose the target, with the individual news frame condition generating more security concerns than people receiving the group frame, the difference was not statistically sig-nifi cant (contrast estimate = .341; s.e. = .269; p = .207). Thus, while the ef-fects of cause predispositions on security concerns were more polarized among those receiving the individual frame than the group frame, this effect was not symmetrical. The interaction is concentrated among those predisposed to support the cause, providing some support for H3a.

Tolerance Judgments

A separate ANCOVA tested the effects of these variables on tolerance of the targeted group after adding security concerns to the list of covari-ates (see Appendix B for full table). Even with the effects of covariates removed, the main effects of cause predisposition and ideology were found to be signifi cant predictors of tolerance toward the targeted group. People were most likely to be tolerant when they encountered the story about activists for a cause they supported, F(1, 572) = 22.10, p < .001, ηp

2 = .04, and when they were more liberal, F(1, 572) = 14.28, p < .001, η2 = .02 (see Table 1). Thus, H1b and H2b received support.

H3b and H4b predicted that the news frame that subjects encountered would moderate the effects of cause predisposition and political ideol-ogy, respectively, on tolerance toward the target, with individual frames amplifying the effects of these dispositional factors. As expected, cause predisposition had a signifi cant interactive effect with the news frames, F(1, 572) = 8.61, p < .01, η2 = .027. Likewise, ideology and news frames

Gil de

Zuniga

- Edu

catio

nal O

nly


were found to interact to shape tolerance toward the targeted-group, F(1, 572) = 3.69, p < .05, ηp

2 = .01. These fi ndings provide some preliminary support for H3b and H4b. To decompose these interactions, we ran ad-ditional contrast tests that parallel the analyses for security concerns.

We fi rst tested whether there was a signifi cant difference in the group tolerance of respondents in the predisposed-to-support and the predis-posed-to-oppose conditions within news frame conditions. This differ-ence was statistically signifi cant among respondents receiving the indi-vidual news frame (contrast estimate = 1.034; s.e. = .192; p = .000) but not among those receiving the group news frame (contrast estimate = -.249; s.e. = .198; p = .210). This set of fi ndings confi rms that it is under the indi-vidual frame condition that the effects of cause predispositions on group tolerance are heightened, whereas these same differences are lessened under the group frame condition. In addition, among those who were predisposed to support the target, respondents receiving the individual frame condition displayed greater tolerance toward the target than peo-ple receiving the group frame condition—even though this relationship fell short of the threshold for statistical signifi cance (contrast estimate = .353; s.e. = .194; p = .069). This difference, however, was signifi cant and in the expected direction among those predisposed to oppose the target (contrast estimate = .432; s.e. = .193; p = .026). Thus, while the effects of cause predispositions on group tolerance were more polarized among those receiving the individual frame than the group frame, once again this effect was not perfectly symmetrical. That is, the difference between frame conditions was statistically signifi cant only among those predis-posed to oppose the cause, providing some support for H3b.

We then tested whether there was a signifi cant difference in tolerance for the target within news frame conditions between respondents with a liberal ideology and those with a more conservative outlook. Once more, this difference was statistically signifi cant among respondents receiving the individual news frame (contrast estimate = 801; s.e. = .204; p = .000) but not among those receiving the group news frame (contrast estimate = -.285; s.e. = .209; p = .173). Thus, it is under the individual frame con-dition that the effects of ideology on group tolerance are heightened, whereas these same differences are reduced to nonsignifi cance under the group frame condition. Contrast tests comparing levels of tolerance between frame conditions, however, did not produce meaningful differ-ences. Among political liberals, respondents encountering individual news frames were not signifi cantly different in their levels of tolerance than respondents encountering group frames (contrast estimate = .128; s.e. = .172; p = .458). The same was true among more conservative respon-dents, with differences between those encountering individual frames and group frames in the expected direction but short of the threshold for statistical signifi cance (contrast estimate = -.388; s.e. = .229; p = .091).

Gil de

Zuniga

- Edu

catio

nal O

nly


Thus, the effects of ideology on group tolerance may be heightened among those receiving the individual frame, but the differences be-tween frame conditions were not statistically signifi cant, providing limited support for H4b.

In sum, our results indicate that news frames do condition the effects of cause predispositions to shape evaluations of security concerns and tolerance judgments. When compared to the group frame, the individual frame polarized security concerns and tolerance judgments according to support for or opposition to a cause. Individual ideology does not have a signifi cant interactive effect with news frames on security concerns; however, a similar interaction with cause predisposition was found for social tolerance. Contrast testing yielded insight into the nature of these interactions and allowed closer examination of hypothesized effects. These results generally provide support for H1a, H1b, H2a, H2b, H3a, and H3b, with limited support for H4b. H4a, however, did not receive empirical support.

DISCUSSION

These fi ndings suggest that story frames and individual political dis-positions, particularly cause predispositions, work jointly to infl uence public reactions toward political radicals and restrictions of their civil liberties. At the most basic level, the results of this study indicate that when respondents were exposed to an extremist activist that they were predisposed to support, they expressed fewer security concerns and greater support for civil liberties. Similar patterns were observed for political ideology, with liberals more tolerant and conservatives less so. That is, when confronted with news about surveillance and profi ling of political radicals, political liberals and cause supporters back Constitu-tional rights. This is not particularly surprising in light of past research (Lipset & Raab, 1970; Sullivan et al., 1981).

What is new and most intriguing here is that individual framing am-plifi es these effects, whereas group framing attenuates them. As a result, individual framing leads to the polarization of opinions about security and civil liberties between cause supporters and cause opponents, with these two groups of respondents making opposing judgments about the same targets of civil liberties restrictions. Likewise, individual story frames made conservatives somewhat more intolerant and liberals some-what more tolerant than they would have been had they encountered the group frame. These fi ndings are particularly noteworthy in light of the dynamic online design of our study, which gauged political preferences in advance of presenting respondents with the radical targets they would be inclined to support or oppose. As a result, the groups for which some

Gil de

Zuniga

- Edu

catio

nal O

nly


respondents favored restrictions were the same ones for which others favored protection, potentially exacerbating the confl ict that this polar-ization of attitudes may produce.

This pattern of results highlights inherent biases in cognitive process-ing of information about individuals and groups. As social psychologists have noted, the perceived entitativity of social units, the degree to which they are understood as being cohesive and consistent, infl uences the ex-tremity of social judgments (Hamilton & Sherman, 1996). Individuals are thought to have greater consistency between attitudes and behaviors be-cause they are self-contained; thus, they are less likely to be constrained from acting on their beliefs, leading to more extreme judgments. The op-posite is true of groups, mitigating security concerns and social tolerance judgments.

Implications for Communication Research

The journalistic tendency to focus on individuals in the construction of news stories appears to encourage audience members to make faster, more confi dent, and more extreme trait inferences (Hamilton & Sher-man, 1996; Susskind et al., 1999). This propensity may provide a broader explanation for Iyengar’s (1991) provisional fi ndings concerning attribu-tions of responsibility for solving social problems. From our perspective, more extreme judgments of episodically—or individually—framed con-tent produces less sympathetic responses about providing for the needy. In the case of civil liberties issues, the tendency of individual framing to amplify citizens’ deeper convictions, both their cause predispositions and political ideologies, may contribute to the polarization of public opinion around “fi ghting terrorism.”

Given the classic liberal emphasis on the rights-bearing individual, news stories will frame civil liberties controversies around the individu-als under government scrutiny, leading audience members who are so disposed to become even more likely to stand up for activists’ free ex-pression, privacy, and autonomy. People who oppose the activists’ cause, however, should be that much less likely to extend these democratic rights when the target of government inquiries is an individual. As a re-sult, individual framing of PATRIOT Act investigations and prosecutions may deepen the schism over civil liberties—with those who are predis-posed to support the cause least likely to sense security concerns and most steadfast in sustaining Constitutional rights while those who oppose the cause holding equally strong, yet sharply divergent positions.

This observation is not just some experimental contrivance, but an ac-tual feature of the current political landscape given the application of the PATRIOT Act. Recent reports confi rm that the Bush administration has used the PATRIOT Act to investigate, infi ltrate, and surreptitiously

Gil de

Zuniga

- Edu

catio

nal O

nly


monitor the activities of antiwar protesters (Goldberg, 2004). Similar to the COINTELPRO programs of the 1970s, which empowered the FBI to engage in surveillance and sabotage of political dissidents, these investi-gations have fostered the sentiment that “civil disobedience, seen during peaceful times as the honorable legacy of heroes like Gandhi and Martin Lu-ther King Jr., is being treated as terrorism’s cousin” (Goldberg, 2004, p. 2).

If the results of this study are any indication, political predispositions such as support for the war may have largely shaped whether audience members saw the targets of government surveillance as participants in democracy or threats to it. More important, however, individual framing of this news (i.e., the norm for such coverage) likely polarized opinion about these targets; evaluations that would have been tempered had the activists been presented as a group. Given the journalistic tendency to personalize news, the clash of values between those supporting the ad-ministration and those opposing their efforts was probably substantially sharpened, diminishing the likelihood of thoughtful debate and delib-eration. In more general terms, the polarizing effects of the individual frames may explain why disagreements over values and ideology in-creasingly defi ne our political culture.

In addition to this insight into the polarization of contemporary po-litical discourse, a theoretical contribution of this study is to advance research on news framing by exploring the moderating infl uences of frames on other variables. Even though we did not fi nd the direct ef-fects of story frames on safety concerns and tolerance judgments, the frames produced conditional effects in relation to group affect and politi-cal ideology. This study extends recent framing research on the interac-tions among frame elements and individual predispositions (Shah, 2001; Shah et al., 2004). We advanced this line of inquiry by examining frames as moderators in the infl uences of core beliefs (i.e., cause predispositions and ideology) that would not be apparent without attention to the con-ditional nature of these news infl uences.

Finally, because of the recent focus on national security, Americans have heightened concerns about public safety. Like tolerance judgments, these concerns could be exacerbated or downplayed by framing issues in different ways. Hinting at the perils of exacerbated effects, many left-ist critics have expressed fears that security concerns are being used to manipulate the public. If one believes that tolerant moderation is desir-able in judgments about extending civil liberties to disliked groups, then the press should make an effort to focus less on individuals and more on groups when covering these confl icts. Such a balance could foster the preservation of civil liberties from unjust or inappropriate infringe-ments. It could also temper the threats people perceive in politically or culturally different groups, as well as increase general tolerance among the American public. The classic liberal marketplace of ideas may always

Gil de

Zuniga

- Edu

catio

nal O

nly


remain an unattainable in the realities of political power struggle; nev-ertheless, journalism still plays a key role in determining how close we might get to that ideal.

APPENDIX A

Question Wording

Age

What is your Age? [ ]

Gender

What is your Sex? 1. Male 0. female

Income

I would like an estimate of your total 2001 household income. Please estimate the com-bined income for all household members from all sources. Was your household income in 2001:

1. Less than $10,0002. Between $10,000 and $30,0003. Between $30,000 and $50,0004. Between $50,000 and $70,0005. Between $70,000 and $90,0006. More than $90,0007. Don’t Know

Education

How many years of school have you completed? (We want a number here, not a degree. Remember to include elementary school). [ ]

Security Concerns

In thinking about the issues raised in the news story, how concerned are you about . . . a. Your own safety from violent acts by extremists groups.b. The safety of other people in the United States from violent acts by extremists

groups.

Tolerance for the Targeted Group

Here is a list of statements that people have made regarding treatment of the [targeted group name]. Please indicate how you feel about each of these statements.

Members of this group should be . . . a. Allowed to work as a teacher in public schoolsb. Allowed to hold public rallies c. Allowed to broadcast public access cable programsd. Prevented from sharing their views over the Internet

Gil de

Zuniga

- Edu

catio

nal O

nly


Cause Support

Below is a list of special interest groups that are politically active in the United States. Toward which of these groups do you feel most negative?

a. The Christian Coalition, a group that seeks to bring a faith-based agenda to politicsb. Greenpeace, a group that seeks to prevent environmental degradationc. National Rifl e Association, a group that seeks to protect the rights of gun ownersd. National Right to Life, a group that seeks to end the practice of abortione. Outproud, a group that seeks to advance the rights of gay and lesbian individuals f. People for the Ethical Treatment of Animals, a group that seeks to promote the hu-

mane treatment of animals. Now, toward which of these groups do you feel most positive?

Ideology

Now thinking in terms of social issues and people’s behavior, would you say you are:1. Very liberal2. Liberal3. Somewhat liberal4. Moderate5. Somewhat conservative6. Conservative7. Very conservative

Maternalism

Here is a list of statements that people have made about their social interactions and trust in various groups. Tell us how you feel about each statement:

a. I worry about what happens to other people, even total strangers. b. It bothers me greatly to see other people get hurt.

Interpersonal Trust


a. If they got the chance, most people would try to take advantage of you.b. Most people are honest.

Interpersonal Discussion


a. I often talk to my family members about politics and current affairs.b. I often talk to my neighbors about politics and current affairs.c. I often talk to my friends about politics and current affairs.d. I often talk to my co-workers about politics and current affairs.

Newspaper Hard News Use

How often do you consume each of the following types of media content?a. Newspaper stories about national politics and international affairsb. Newspaper stories about local government and community issuesApart from how often you consume these types of content, how much attention do you

pay to them? How much attention do you pay to:

Gil de

Zuniga

- Edu

catio

nal O

nly


a. Newspaper stories about national politics and international affairsb. Newspaper stories about local government and community issues

TV Hard News Use

How often do you consume each of the following types of media content?a. National television news programsb. Local television news programsApart from how often you consume these types of content, how much attention do you

pay to them? How much attention do you pay to:a. Television stories about national politics and international affairsb. Television stories about local government and community issues

APPENDIX B

Covariates

Gender F(1, 576) = 6.78, p = .01, ηp2 = .01

Age F(1, 576) = 3.52, p = .06, ηp2 = .01

Education F(1, 576) = .00, p = .99, ηp2 = .00

Income F(1, 576) = .17, p = .68, ηp2 = .00

Paternalism F(1, 576) = 13.57, p = .00, ηp2 = .02

Interpersonal trust F(1, 576) = 16.74, p = .00, ηp2 = .03

Newspaper hard news use F(1, 576) = .17, p = .68, ηp2 = .00

Television hard news use F(1, 576) = 16.34, p = .00, ηp2 = .03

Interpersonal discussion F(1, 576) = 1.11, p = .29, ηp2 = .00

Source cue manipulation F(1, 576) = .02, p = .89, ηp2 = .00

Number of visited manipulation pages F(1, 576) = 1.97, p = .16, ηp2 = .00

Main Effects

Cause predisposition F(1, 576) = 11.76, p = .00, ηp2 = .02

Political ideology F(1, 576) = 6.03, p = .01, ηp2 = .01

Frame manipulation F(1, 576) = .19, p = .67, ηp2 = .00

Interactions

Frame * Predisposition F(1, 576) = 5.23, p = .02, ηp2 = .01

Frame * Ideology F(1, 576) = .12, p = .73, ηp2 = .00

ANCOVAs with the Main Effects, Interactions, and Covariates for Security Concerns

Security concerns

Gil de

Zuniga

- Edu

catio

nal O

nly


ANCOVAs with the Main Effects, Interactions, and Covariates for Group Tolerance

Group tolerance

Covariates

Gender F(1, 572) = 2.57, p = .11, ηp2 = .00

Age F(1, 572) = .55, p = .46, ηp2 = .00

Education F(1, 572) = 3.12, p = .08, ηp2 = .01

Income F(1, 572) = .78, p = .38, ηp2 = .00

Paternalism F(1, 572) = 2.56, p = .11, ηp2 = .00

Interpersonal trust F(1, 572) = 4.25, p = .04, ηp2 = .01

Newspaper hard news use F(1, 572) = 8.80, p = .00, ηp2 = .02

Television hard news use F(1, 572) = 6.96, p = .01, ηp2 = .01

Interpersonal discussion F(1, 572) = 5.94, p = .02, ηp2 = .01

Source cue manipulation F(1, 572) = .39, p = .53, ηp2 = .00

Number of visited manipulation pages F(1, 572) = 2.73, p = .10, ηp2 = .01

Security concerns F(1, 572) = 39.25, p = .00, ηp2 = .06

Main Effects

Cause predisposition F(1, 572) = 22.10, p = .00, ηp2 = .04

Political ideology F(1, 572) = 14.28, p = .00, ηp2 = .02

Frame manipulation F(1, 572) = .60, p = .44, ηp2 = .00

Interactions

Frame * Predisposition F(1, 572) = 8.61, p = .00, ηp2 = .02

Frame * Ideology F(1, 572) = 3.69, p = .05, ηp2 = .01

NOTES

1. Here and throughout the manuscript we refer to individual frames as amplifying the effects of cause predispositions and groups frames as attenuating the effects of cause predispositions. We recognize that there is no formal control group for the news framing manipulations. Of course, to create a news story without an orienting frame would be impossible, as would asking respondents questions about a hypothetical target without a news story presenting the target as some social level—either individual or group. As such, we use the terms “amplify” and “attenuate” as linguistic conveniences rather than specifi c theoretical claims.

2. Many participants were legitimately unable to participate because they lacked in-ternet access; therefore, we report this fi gure instead of a traditionally calculated response rate. Even given this more generous method of calculation, fi nal participation rates are

Gil de

Zuniga

- Edu

catio

nal O

nly


quite low, and the results provided here should be viewed in terms of experimental fi nd-ings that show strong internal validity, rather than as an externally valid assessment of the feelings of the Dane County, Wisconsin community.

3. The names of the six groups and their fi ctional counterparts: (a) the Christian Coali-tion–United Christian Front, (b) Greenpeace–Earth Defense Front, (c) National Rifl e As-sociation–Arm America Front, (d) National Right to Life–Unborn Defense Front, (e) Out-proud–United Queer Front, and (f) People for the Ethical Treatment of Animals–Animal Liberation Front.

4. Below the accepted threshold of .70, the Cronbach’s alpha yielded by the interper-sonal trust index nonetheless was not regarded as a threat to the validity of the measure. As Carmines and Zeller (1979) point out, a two-item scale with an interitem correlation of .4 would fall shy of standards for alpha reliability (.57), but a six-item index with the same average interitem correlation would easily exceed standards (.80). We argue that a robust association alongside clear face validity implies that items are effectively measuring an underlying concept.

5. For the ANCOVA, we created a dummy variable for ideology by dichotomizing the continuous variable (Liberal = 1; Conservative = 0). We combined moderate and conserva-tive respondents into one group labeled “conservative” when creating this dummy vari-able. This was done to avoid creating vastly different case counts in the dummy variable because 64.5% of the total sample was liberal.

6. A power analysis that was conducted for the main and the interactive effects assum-ing two-tailed alpha of .5 revealed powers of .93, .69 and .63 for small effect size (respec-tively 0.2, 0.1 and .01) and 0.6 for no effect (0.0).

7. A power analysis that was conducted for the main and the interactive effects assum-ing two-tailed alpha of .5 revealed a power of 1.00 for medium effect size (0.4) and powers of .97, .83 and .48 for small effect size (respectively 0.2, 0.2 and 0.1).

REFERENCES

Ball-Rokeach, S. J., & Loges, W. E. (1994). Choosing equality: The correspondence between attitudes about race and the value of equality. Journal of Social Issues, 50, 9–19.

Ball-Rokeach, S. J., & Loges, W. E. (1996). Making choices: Media roles in the construction of value choices. In C. Seligman & M. Zanna (Eds.), The psychology of values: The Ontario symposium (Vol. 8, pp. 277–298). Mahwah, NJ: Lawrence Erlbaum.

Billig, M., & Tajfel, H. (1973). Social categorization and similarity in intergroup behavior. European Journal of Social Psychology, 3, 27–51.

Bobo, L. & Licari, F. (1989). Education and political tolerance: Testing the effects of cogni-tive sophistication and target group affect. Public Opinion Quarterly, 53, 285–308.

Carmines, E. G., & Zeller, R. A. (1979). Reliability and validity assessment. Newbury Park, CA: Sage

Chong, D. (1994). Tolerance and social adjustment to new norms and practices. Political Behavior, 16, 21–53.

Gaertner, L., & Insko, C.A. (2001). On the measurement of social orientations in the mini-mal group paradigm: Norms as moderators of the expression of intergroup bias. Euro-pean Journal of Social Psychology, 31, 143–154.

Goldberg, M. (2004).Lost liberties: Outlawing dissent. Salon. Retrieved from http://www.salonmag.com/news/feature/2004/02/11/cointelpro/index_np.html

Golebiowska, E. A. (1996). The “pictures in our heads” and individual-targeted tolerance. The Journal of Politics, 58(4), 1010–1034.

Green, J. C., & Guth, J. L. (1991). Religion, representatives, and roll calls. Legislative Studies Quarterly, 16, 571–584.

Gil de

Zuniga

- Edu

catio

nal O

nly


Hamilton, D. L., & Sherman, S. J. (1996). Perceiving persons and groups. Psychological Re-view, 103(2), 336–355.

Hurwitz, J., & Mondak, J. J. (2002). Democratic principles, discrimination and political in-tolerance. British Journal of Political Science, 32, 93–118.

Iyengar, S. (1991). Is anyone responsible? How television frames political issues. Chicago: Uni-versity of Chicago Press.

Keppel, G. (1991). Design and analysis: A researcher’s handbook. Englewood Cliffs, NJ: Pren-tice Hall.

Kinder, D. R., & Sears, D. O. (1985). Public opinion and political action. In G. Lindzey & E. Aronson (Eds.), The handbook of social psychology (pp. 659–741). New York: Random House.

Kirk, R. E. (1995). Experimental design: Procedures for the behavioral sciences. Pacifi c Grove, CA: Brooks/Cole Publishing Company.

Krosnick, J. A., Berent, M. K., & Boninger, D. S. (1994). Pockets of responsibility in the American electorate: Findings of a research program on attitude importance. Political Communication, 11, 391–411.

Kuklinski, J. H., Riggle, E., Ottati, V. Schwarz, N., & Wyer, R. S. Jr. (1991). The cognitive and affective bases of political tolerance judgments. American Journal of Political Science, 35(1), 1–27.

Lakoff, G. (2004). Don’t think of an elephant: Know your values and frame the debate, White River Junction, VT: Chelsea Green Publishing Company

Lipset, S. M., & Raab, E. (1970). The politics of unreason: Right-wing extremism in America, 1790–1970. New York: Harper and Row.

Marcus, G. E., Sullivan, J. L., Theiss-Morse, E., & Wood, S., L. (1995). With malice towards some: How people make civil liberties judgments. New York: Cambridge University Press.

McLeod, D.M., & Hertog, J.K. (1999). Social control and the mass media’s role in the regulation of protest groups: The communicative acts perspective. In D. Demers and K. Viswanath (Eds.) Mass media, social control and social change. Ames: Iowa State Uni-versity Press.

Nelson, T., Clawson, R., & Oxley, Z. (1997). Media framing of a civil liberties confl ict and its effect on tolerance. The American Political Science Review, 91, 567–583.

Ng, S. H. (1981). Equity theory and the allocation of rewards between groups. European Journal of Social Psychology, 11, 439–444.

Nunn, C. Z., Crockett, H. J., & Williams, J. A. (1978). Tolerance for nonconformity: A national survey of Americans’ changing commitment to civil liberties. San Francisco: Jossey-Bass Publishers.

Price, V., & Tewksbury, D. (1997). News values and public opinion: A theoretical account of media priming and framing. In G. Barnett & F. Boster (Eds.) Progress in communication sciences (pp.173–212). Norwood, NJ: Ablex.

Rabbie, J. M., Schot, J. C., & Visser, L. (1989). Social identity theory: A conceptual and em-pirical critique from the perspective of a behavioral interaction model. European Journal of Social Psychology, 19, 171–202.

Shah, D. V. (2001). The collision of convictions: Values framing and value judgments. In R. P. Hart and D. Shaw (Eds.), Communication and U.S. elections: New agendas. Lanham, MD: Rowman and Littlefi eld.

Shah, D. V., Domke, D., & Wackman, D. B. (1996). To thine own self be true: Values, framing and voter decision-making strategies. Communication Research, 23, 509–560.

Shah, D. V., Kwak, N., Schmierbach, M., Zubric, J. (2004). The interplay of news frames on cognitive complexity. Human Communication Research, 30, 102–120.

Sotirovic, M., & McLeod, J. M. (2001). Values, communication behavior, and political par-ticipation. Political Communication, 18, 273–300.

Stouffer, S. A., (1963). Communism, conformity and civil liberties: A cross-section of the nation speaks its mind. Gloucester, MA: Peter Smith. Original work published 1955

Gil de

Zuniga

- Edu

catio

nal O

nly


Sullivan, J. L., & Marcus, G. E. (1988). A note on “trends in political tolerance.” Public Opin-ion Quarterly, 52, 26–32.

Sullivan, J. L., Marcus, G. E., Feldman, S., & Piereson, J. (1981). The sources of political toler-ance: A multivariate analysis. The American Political Science Review, 75, 92–106.

Sullivan, J. L., Piereson J., & Marcus, G. E. (1979). An Alternative conceptualization of po-litical tolerance: Illusory increases 1950s-1970s. The American Political Science Review, 73, 781–794.

Sullivan, J. L., Piereson, J., & Marcus, G. E. (1982). Political tolerance and American democracy. Chicago: University of Chicago Press.

Susskind, J., Maurer, K., Thakkar, V., Hamilton, D.L., & Sherman, J.W. (1999). Perceiving individuals and groups: Expectancies, dispositional inferences, and causal attributions. Journal of Personality and Social Psychology, 76(2), 181–191.

Tajfel, H., & Billig, M. (1974). Familiarity and categorization in intergroup behavior. Journal of Experimental Social Psychology, 10, 159–170.

Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup confl ict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp.33–47).

Tajfel, H., Billig, M., Bundy, R. P., & Flarnent, C. (1971). Social categorization and intergroup behavior. European Journal of Social Psychology, 1, 149–178.

Tetlock, P. E. (1986). A value pluralism model of ideological reasoning. Journal of Personality and Social Psychology, 50, 819–827.

Tetlock, P. E., Armor, D., & Peterson, R. S. (1994). The slavery debate in Antebellum Amer-ica: Cognitive Style, value confl ict, and the limits of compromise, Journal of Personality and Social Psychology, 66, 115–126.

Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S. (1987). Rediscovering the social group: A self-categorization theory. New York: Basil Blackwell.

Gil de

Zuniga

- Edu

catio

nal O

nly

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Personifying the radical: How news framing polarizes security concerns and tolerance judgments.

Documents